diff options
Diffstat (limited to 'doc/limits.html')
| -rw-r--r-- | doc/limits.html | 49 |
1 files changed, 49 insertions, 0 deletions
diff --git a/doc/limits.html b/doc/limits.html new file mode 100644 index 0000000..d09fe5b --- /dev/null +++ b/doc/limits.html @@ -0,0 +1,49 @@ + +<!DOCTYPE html> + +<html lang="en"> + <head> + <meta charset="utf-8"> + <title>Limitations - Packrat Docs</title> + <link href="default.css" rel="stylesheet"> + </head> + + <body> + + + <h2>Limitations</h2> + + <a href="index.html">Return to Contents</a> + + + <h4>Recognisers</h4> + <p>While this package can be used to construct recognisers easily, this would likely + not be very efficient compared to a library more specialised for that purpose.</p> + + <h4>Cyclic Grammars</h4> + <p>Grammars that have cycles in them will be treated as any other form of left recursion + and eventually be curtailed. This means that the resulting parse graph will not have + any cycles. If you want a cyclic grammar to result in a cyclic result, then this library + is not what you are after.</p> + + <h4>Generics</h4> + <p>This library makes very heavy use of generics. If your Ada compiler has any issues + handling generics then you may experience suboptimal performance.</p> + + <h4>Parse Graph Data Structure</h4> + <p>In order to keep the parse graph to polynomial space complexity, it is a specialised + data structure and is not the same thing as a more standard directed graph.</p> + + <h4>Piecewise Parsing</h4> + <p>Due to the possibility of incomplete parses, in the worst case using piecewise parsing + can result in the lexer or parser holding the entire input in memory before returning a + result.</p> + + <h4>Error Recovery</h4> + <p>While error messages can be informative, there is currently no consideration given to + being able to recover from an error and continue parsing.</p> + + + </body> +</html> + |